stable-diffusion-bot

Discordで話しかけるとStable Diffusionくんが絵を生成してくれるよ

howto

server:

$ cd server
$ poetry install
$ WORKER_SECRET='...' DISCORD_TOKEN='...' poetry run python diffusive_phone_server/main.py

worker:

$ cd worker
$ poetry install
$ SERVER_SECRET='...' DIFFUSION_TOKEN='...' poetry run python diffusive_phone_worker/main.py

TODO

• switch to using Discord-bot-based communication

• implement user authentication

• add progress indicator?

• generate multiple images at once?

• use a worker queue for generating multiple images in parallel?

• add zstd compression for websocket comms

• send message notifying user if cannot find worker

• deploy to AWS

  • domain name actually isn't really necessary...

• implement game

  • it would be nice if we could use buttons...

• implement control channel commands